A simple method of inferring pairwise gene interactions from microarray time series data

Microarray data provide a rich resource for analysis of gene expression. Inferring the gene-gene relationships from microarray data is one of the most burgeoning research topics during recent years. While a great many methods have been widely applied on microarray data to discover genes with similar expression patterns, this paper proposes a new simple method to infer the pairwise gene interactions. We have experimented our method on an open microarray time series data with 24 time points of 20 genes. Experimental results show that our method can find not only some relationships that are already found in other literatures, but also can reveal some previously unknown interactions. Furthermore, compared with BNs, our method is very simple and easy to implement, whereas it has nearly the same ability as BNs in terms of the interactions inference.